From: David J. <dav...@di...> - 2002-07-03 15:25:22
|
Those reply to headers got me again. On 2002.07.03 08:40:21 -0400 David Jencks wrote: My experience is that, whenever I write them, my software only does what I think if there is a unit test. Although the tests are not terribly complete, I think the client-java driver would not have gotten very far without the junit test framework and tests. My experience is that after the testing framework is set up it is very easy to write new tests, and gets to be almost fun quickly;-) There appear to be several c++ unit test frameworks now: http://www.xprogramming.com/software.htm I can't really judge which one would be most appropriate, but based on my java experience we will get great rewards if someone takes the time to set up the infrastructure. Thanks david jencks On 2002.07.03 07:04:13 -0400 Pavel Cisar wrote: > Hi all, > > Developers often use home-made test applications to test basic > functionality of parts they're working on. They are not always developed > nor complete, but if they are developed, they are not deleted afterwards > and could be reused in test system or by QA group. They don't need to be > so extreme like in eXtreme programming, and developed *before* any real > development take place (although XP approach would be nice :), but *any* > such application has its value. > > Pro's and Con's: > > This technique impose significant burden to core developers, if they're > required to write test cases along their code. Of course, these tests > could be written by others, but this task require intimate knowledge at > least tested piece of source code. Tests must be written in certain > standard way, to assure seamless integration into automatic test system. > Test applications that would test correctness of various low-level > functions or modules/subsystems can quickly uncover bugs that are > otherwise hard to catch at higher level, especially those that are > related to parameter values and execution flow handling (if/else/switch). > > From my experience, around 50% of all bugs are related to bad > if/else/switch handling (undetected/badly handled values etc.). These > tests can also work as a specification for produced code. > > Implementation: > > We can initiate a sub-project in Firebird to build such test programs for > > selected libraries and separate subsystems (like memory management, > optimizer, parser etc.) and/or declare a policy for core developers > about when and how write such test cases. The basic prerequisite to start > > it successfully is a good source code documentation (file dependencies, > list of routines and symbols etc.) to allow a seasoned C/C++ developers > to get on track quickly. We have at (least basic) such documentation, and > > for rest it's exactly where GLOBAL tagging systems would come handy. > > Current status: > > We don't have anything like this for Firebird right now. > > Questions for you: > 1) Do you think that we should use low-level test cases ? > > > 2) If yes, what parts of core engine do you suggest to be a subject of > this testing ? > > > 3) If yes, what procedure and/or toolkit do you recommend ? > > > Your comments would be greatly appreciated. > Best regards > Pavel Cisar > http://www.ibphoenix.com > For all your upto date Firebird and > InterBase information > > > > ------------------------------------------------------- > This sf.net email is sponsored by:ThinkGeek > No, I will not fix your computer. > http://thinkgeek.com/sf > _______________________________________________ > Firebird-test mailing list > Fir...@li... > https://lists.sourceforge.net/lists/listinfo/firebird-test > > |